mcp-server-bwt

mcp-server-bwt

MCP server for Bing Webmaster Tools

This MCP (Model Context Protocol) server provides a bridge between supported AI assistants like Claude or Cursor and the Bing Webmaster Tools API. It exposes all Bing Webmaster Tools functionality available via bing-webmaster-tools as MCP tools that can be used by AI assistants to interact with your Bing Webmaster Tools account.

Example Usage with Claude

Once configured, you can use the MCP server with Claude to interact with your Bing Webmaster Tools account. Here are some example prompts:

• "List all my verified sites in Bing Webmaster Tools"
• "Submit my homepage for indexing"
• "Get traffic statistics for my website"
• "Check for any crawling issues on my site"
• "Get keyword statistics for 'my product'"

Claude will use the appropriate MCP tools to fulfill your requests.

Requirements

• uv, which provides uvx and downloads a suitable Python for you
• Bing Webmaster Tools API Key

Installation

mcp-server-bwt is published on PyPI. Your MCP client runs it with uvx, so you don't need a separate install step. Every config below passes your API key through BING_WEBMASTER_API_KEY.

Claude Desktop

In your Claude config specify:

"mcpServers": {
  "mcp_server_bwt": {
    "command": "uvx",
    "args": ["mcp-server-bwt"],
    "env": {
      "BING_WEBMASTER_API_KEY": "YOUR_API_KEY_HERE"
    }
  }
}

Claude Code

claude mcp add bwt -e BING_WEBMASTER_API_KEY=YOUR_API_KEY_HERE -- uvx mcp-server-bwt

Cursor

In .cursor/mcp.json (per project) or ~/.cursor/mcp.json (global) add:

"mcpServers": {
  "bwt": {
    "command": "uvx",
    "args": ["mcp-server-bwt"],
    "env": {
      "BING_WEBMASTER_API_KEY": "YOUR_API_KEY_HERE"
    }
  }
}

Zed

In your Zed settings.json add:

"context_servers": {
  "bwtServer": {
    "command": "uvx",
    "args": ["mcp-server-bwt"],
    "env": {
      "BING_WEBMASTER_API_KEY": "YOUR_API_KEY_HERE"
    }
  }
}

Versions and upgrades

uvx caches the version it downloaded first and keeps reusing it. To pick up a new release, use "args": ["mcp-server-bwt@latest"], or run uv cache clean mcp-server-bwt once. To pin a version, use "args": ["[email protected]"]. Release notes are in CHANGELOG.md.

Available Tools

The server provides the following Bing Webmaster Tools API functionality (more info in the API docs):

Paged results

Tools that return a list (for example get_query_stats, get_page_stats and get_crawl_issues) return at most 50 rows per call, so a large site's history doesn't overflow the client's context window. They take two optional parameters:

• limit: rows per call (1–500, default 50; see BING_WEBMASTER_PAGE_SIZE below)
• offset: index of the first row (default 0)

Each paged result ends with a line such as Showing 50 of 1200 rows from offset 0; next_offset=50. The same values are in the result's _meta under mcp-server-bwt/pagination (total, offset, limit, next_offset). To get the next page, call again with offset set to next_offset.

The server keeps each list it downloads for 5 minutes (see BING_WEBMASTER_CACHE_TTL below). Repeat calls with the same arguments, including later pages, are served from that copy instead of downloading the whole list again. A tool that changes data, such as add_blocked_url, clears the cached lists of its own area (here the blocking tools), so the next read shows the change. Site tools such as remove_site clear every cached list. Changes made elsewhere, for example in the Bing web interface, can take up to 5 minutes to appear. Stats that Bing computes, such as traffic, crawl and keyword data, can be up to 5 minutes old.

Migrating from 0.1.x: list tools used to return every row. To get that behavior back, set BING_WEBMASTER_PAGE_SIZE=0 in the server's env. To also fetch fresh data on every call, set BING_WEBMASTER_CACHE_TTL=0.

Site Management

• get_sites: List all verified sites in your Bing Webmaster Tools account
• add_site: Add a new site to your account
• verify_site: Verify ownership of a site
• remove_site: Remove a site from your account
• get_site_roles: Get roles for a specific site
• add_site_roles: Add roles to a site
• remove_site_role: Remove a role from a site
• get_site_moves: Get information about site moves
• submit_site_move: Submit a site move request

URL Submission

• submit_url: Submit a single URL for indexing
• submit_url_batch: Submit multiple URLs for indexing in a batch
• submit_content: Submit content for indexing
• submit_feed: Submit a feed for indexing
• get_feeds: Get all submitted feeds
• get_feed_details: Get details about a specific feed
• remove_feed: Remove a feed from your account
• get_url_submission_quota: Check your URL submission quota
• get_content_submission_quota: Check your content submission quota
• fetch_url: Fetch a URL for indexing
• get_fetched_urls: Get all fetched URLs
• get_fetched_url_details: Get details about a specific fetched URL

Traffic Analysis

• get_query_stats: Get statistics for search queries
• get_query_traffic_stats: Get traffic statistics for search queries
• get_query_page_stats: Get page statistics for search queries
• get_query_page_detail_stats: Get detailed page statistics for search queries
• get_page_stats: Get statistics for pages
• get_page_query_stats: Get query statistics for pages
• get_rank_and_traffic_stats: Get rank and traffic statistics

Crawling

• get_crawl_stats: Get crawling statistics
• get_crawl_settings: Get crawling settings
• save_crawl_settings: Save crawling settings
• get_crawl_issues: Get crawling issues

Keyword Analysis

• get_keyword: Get information about a keyword
• get_keyword_stats: Get statistics for a keyword
• get_related_keywords: Get related keywords

Link Analysis

• get_link_counts: Get link counts
• get_url_links: Get links for a URL
• get_deep_link: Get deep link information
• get_deep_link_blocks: Get deep link blocks
• add_deep_link_block: Add a deep link block
• remove_deep_link_block: Remove a deep link block
• update_deep_link: Update a deep link
• get_deep_link_algo_urls: Get deep link algorithm URLs
• get_connected_pages: Get connected pages
• add_connected_page: Add a connected page

Content Management

• get_url_info: Get information about a URL
• get_url_traffic_info: Get traffic information for a URL
• get_children_url_info: Get information about child URLs
• get_children_url_traffic_info: Get traffic information for child URLs

Content Blocking

• get_blocked_urls: Get blocked URLs
• add_blocked_url: Add a URL to the blocked list
• remove_blocked_url: Remove a URL from the blocked list
• get_active_page_preview_blocks: Get active page preview blocks
• add_page_preview_block: Add a page preview block
• remove_page_preview_block: Remove a page preview block

Regional Settings

• get_country_region_settings: Get country/region settings
• add_country_region_settings: Add country/region settings
• remove_country_region_settings: Remove country/region settings

URL Management

• get_query_parameters: Get query parameters
• add_query_parameter: Add a query parameter
• remove_query_parameter: Remove a query parameter
• enable_disable_query_parameter: Enable or disable a query parameter

Development

Development needs Python >= 3.13 and uv. Node.js is only needed for make mcp_inspector.

make install installs the dependencies and the git hooks in lefthook.yml:

• pre-commit runs ruff check --fix and ruff format on your staged *.py files. Fixes are staged into the same commit. The commit fails only on findings ruff can't fix.
• pre-push runs mypy --strict, plus the test suite when the push includes *.py files and uv lock --check when it includes pyproject.toml or uv.lock.

To skip the hooks once, use git commit --no-verify or git push --no-verify, or set LEFTHOOK=0.

To run all tests:

BING_WEBMASTER_API_KEY=dummy make test

To build the app:

make build

To lint the project:

make lint

To format the project:

make format

Running from a local checkout

To try your changes in a client, point it at the checkout's virtualenv after make install. This is for contributors only. Users should install from PyPI as shown above.

Claude Desktop:

"mcpServers": {
  "bwtServer": {
    "command": "/PATH/TO/mcp-server-bwt/.venv/bin/python",
    "args": ["/PATH/TO/mcp-server-bwt/mcp_server_bwt/main.py"],
    "env": {
      "BING_WEBMASTER_API_KEY": "YOUR_API_KEY_HERE"
    }
  }
}

Zed:

"context_servers": {
  "bwtServer": {
    "command": "/PATH/TO/mcp-server-bwt/.venv/bin/python",
    "args": ["/PATH/TO/mcp-server-bwt/mcp_server_bwt/main.py"],
    "env": {
      "BING_WEBMASTER_API_KEY": "YOUR_API_KEY_HERE"
    }
  }
}

Environment Variables

The following environment variables are required:

• BING_WEBMASTER_API_KEY: Your Bing Webmaster Tools API key

Optional:

• BING_WEBMASTER_PAGE_SIZE: default number of rows list tools return per call (1–500, default 50). 0 turns paging off: list tools return every row with no summary line, unless a caller passes limit or offset, and the 500 cap is lifted. The server doesn't start if the value is invalid.
• BING_WEBMASTER_CACHE_TTL: seconds a list tool keeps a downloaded list for later calls with the same arguments (0–86400, default 300). Results can be this many seconds stale, except that a write tool clears the cached lists of its own area, and a site tool (add_site, remove_site, …) clears them all. 0 turns caching off, so every call reaches the Bing API. The server doesn't start if the value is invalid.

Starting the Server

To start the MCP server:

make start

MCP Inspector

You can use the MCP inspector to test the server:

make mcp_inspector

Creating from Template

This MCP server was created from a cookiecutter template. To create a similar one, run:

uvx cookiecutter gh:zizzfizzix/python-base-mcp-server

License

mcp-server-bwt is licensed under the MIT License. This means you are free to use, modify, and distribute the software, subject to the terms and conditions of the MIT License. For more details, please see the LICENSE file in the project repository.